Abstract This paper traces the history of the notion of gradience in language studies. Gradience is a cover term to designate a spectrum of continuous phenomena in language, from categories at the level of grammar to sounds at the level of phonetics. The focus here is on grammatical gradience. After discussing a number of philosophical ideas which form the backdrop to gradience the paper proceeds to outline a history of ideas on grammatical fuzziness. It concludes by proposing a compromise between the generally opposed ideas on categorization put forward in formal linguistics, and those adopted in cognitive and functional linguistics. This is achieved by applying morphosyntactic tests to linguistic formatives that display syntactic behaviour which is associated with more than one category, so as to establish whether the item in question belongs to one class or another. The morphosyntactic similarity of a category to another category in a particular syntactic configuration is then modelled by appealing to the notion of `convergence'.